Transfer Students in Institutions of Higher Education .NCES 77-250
EDUCATION National Longitudinal Study of High School Seniors
This report was prepared for the National Center for Education Statistics by Research Triangle Institute under contract No. OEC-O-73-66666 with the Education Division. Contractors undertaking such projects under Government sponsorship are encouraged to express freely their professional judgment. This report, therefore, does not necessarily represent positions or policies of the Education Division, and no official endorsement should be inferred.
U.S. DEPARTMENT OF HEALTH, EDUCATION, AND WELFARE Joseph A. Califano, Jr., Secretary
Education Division Mary F. Berry, Assistant Secretary for Education
National Center for Education Statistics Marie D. Eldridge, Administrator
.~ II
FOREW ORD
The National Longitudinal Study of the High School Class of 1972,. a survey initiated by and con- ducted for the National Center for Education Statistics, began in the spring of 1972 with over 1,000 in-school group administrations of survey forms to a sample of approximately, 18,000 seniors. In the followup surveys, the sample included almost 5,000 additional students from sample schools that were unable to participate in the base-year survey.
The data collected from the in-school and two followup surveys have been merged and processed. Results are being presented in a series of reports designed to highlight selected findings in educational, career, and occupational development. This report contains 'information about those students who moved among institutions of higher education over 2 years since initial matriculation. It includes the extent of transfer, the students' reasons for transfer, and variables associated with transfer.
Continuing followup requests for data from these individuals are planned through 1979 and perhaps beyond. This series of repeated observations will permit the examination of the relationships between schooling, work, and other experience to subsequent career choices as well as educational and labor- force participation of each of the selected individuals. Such information and the resultant analyses are limportant to those engaged in formulatinglegislative proposals,and educational policy.
This report was prepared by Samuel S. Peng of the Research Triangle Institute under contract with the U. S. Department of Health, Education, and Welfare for the National Center for Education Statis- tics. The project director was J. P. Bailey, Jr., of RTI's Center for Educational Research and Develop- ment.
Francis V. Corrigan, Deputy Director Elmer F. Collins, Chief Division ofMultilevel Education Statistics Longitudinal Studies Branch
iii
CONTENTS--Continued
Page Figures-
1. Percentages and estimated national totals of students who transferred during or at the end of their first year in college.... I....I................... 6
2. Percentages and estimated national totals of students who transferred during or at the end of their second year in college.......................... 7
3. Total percentages and estimated number of students who transferred over 2 years.. . .... 9 B-i., Flow chart of college entries and transfers (4-year college) .. ............ 67 B-2. Flow chart of college entries and transfers (2-,year college) .68........ ...
Tables 1.' Percentage ofsample members by various background characteristics..... ...... 3 2. Transfer specification.......................8 3-a. Percentages of 4-year college students who transferred by the end of the sophomore year: by
subgroups ............... ............. 11 3-b.~Group differences in transfer rates (in percent) from 4-year colleges ....... ..... 12 4-a. Percentage of 2-year -college students who transferred by the end of the sophomore year: by
subgroups.13............................. 4-b. Group differences in transfer rates (in percent) from 2-year colleges ....... .... 14 5-a. Transfer rate (in percent) by'type ofinstitutional control.. ...... I........ 15 5-l. Transfer students cross-classified by type of control of initial and destination colleges...... 16 6-a. Transfer rates (in percent) of 4-year colleges by selectivity level ofinstitutions.. ...... 17 6-b. Initial and destination college selectivity level of 4-year college transfer students ... .. ... 17 7-a. Transfer rate (in percent) by size of institution.. ........ e ...... 18 7-b. Transfer students cross-classified by size of initial and destination colleges......I... . 18 8. Weighted means and standard deviations for various college-going status groups on background
variables (4-year colleges) .......................... 22 9-a. Test statistics for the comparisonbetween persisters and 4-*4 transfers.. ........ 23 9-b. Test statistics for the comparison between 4-~2 transfers and students who remained in 4-year
colleges (i.e., persisters and 4-*4 transfers) .......... .... I...... 24 9-c. Test statistics for the comparison between 4-*2 transfers and withdrawals.......... 25
10. Weighted means and standard deviations for various college-going status groups on background variables (2-year colleges) .......... I................26
11I-a. Test statistics for the comparison between persisters and 2-~2 transfers........... 27 11I-b. Test statistics for the comparison between 2->4 transfers and students who remained in 2-year
colleges (i.e., persisters and 2-~2 transfers) .. .................. 28 11I -c. Test statistics for the comparison between 2->4 transfers and withdrawals .......... 29 11 A. Test statistics for the comparison between 2-N transfers and graduates..... .29.... 12. Weighted means and common standard deviations of the selected variables for native and 2-+4
transfer students.................... I.........32 13. Test statistics for the differences between native students and 2->4 transfers on the selected
variables.~~~~~~~~~~~~~~~~~32 14-a. Percentage of native and transfer students who received various kinds of scholarships, fellowships,
or grants ............................... 34 14-b. Tests of fit for the logistic model.....I.......... .......... 34
Foreword' ....... ........ .... ........ 1....11
I. Introduction ............... 1.............. A. Background of the Study .1.. ................ B. Purposes of the Study .1.................... ... C. The Data Base..... ........................ 2
DWegting and Significance Testing.2........ E. An Overview of the Remainder of the Report..........4
II. Extent of College Transfers..................... 5 A. Total Transfer Rates ............... .......... B. Transfer Rates by Subgroups Defined by Background Variables. ........... .9
C. Transfer Rates by the Type of Institution.. .................. 15 D. Summary and Discussion .... ....... ........... 19
III. Comparisons Between Transfer and Nontransfer Students ..... .21..... A.4-Yar Colege Trnsfer 22Comarison Betwen and Nntransers;
B. Comparisons Between 2.Year College Transfers and Nontransfers........ 22
C. Summary and.Discussion ............ I....... ... 30 IV. Comparisons Between 2->4 Transfers and 4-Year College Native Students . .... 31
A. Comparisons on Background Variables and Individual Characteristics .. ........ 31 -B. Comparisons on Financial Aid Status ..... .............. 33 C. Comparisons on Academic Performance............ ....... 36, D. Comparisons on the Evaluation of College Education............... 36 E. Summary and Discussion ...... ................. 37
V. Students' Self-Reported Reasons for Changing Schools....I.. ... .... 39 A. Reasons Given by 4->2 Transfers ....... I...........40 B. Reasons Given by 2-4 Transfers .41........ ........... C. Reasons Given by Horizontal Transfers............ . 42.... D. Summary.................. ........... 42
VI. Person-InstitutionIncongruency and Transferring................43 A. Ability-Challenge Incongruency ....................... 43 B. Expectation Incongruency .44.....I... ............ C. Financial Support and Expense Incongruency.46........... D. Summary and Discussion ......... I.............47.
VII. Conclusion and Implications..........49; References ............................... 53 Appendixes A. Description of the NLS Data Base: Sample, Procedures, and Instruments.........63 B. Percentage and Estimated Total at Each Study Status Point, Over Three Points in Time. 67, C.- Percentage of Students in Various:Study-Status Categories Crossed by Background Variables. . ~69 D. Definition of Psychological Constructs: Self-Esteem, Locus of Control, and Life Goals....
15-a. Percentage of native and transfer students who received various kinds ofloans . . ........ 34
16 Percentage of student who indicated receiving various kinds of scholarships, fellowships, or; grants . ..... ..............
15-b. Tests of fit for the logistic model for the~ association of SES, types of students, and receipt of 6.loans............... I..................3
17. Percentage of students who reported receiving various kinds ofloans....35 18.: Distributions of the student self-reported college grade-point averges .3
19. Percentage of students who indicated dissatisfaction with various aspects of colg dctOn -..- 37 20. Reasons freshman transfers gave for changing schools... ....... 39 21. Reasons sophomore transfers gave for changing schools ........ 40....... 22. Percentage of freshmen transfers indicating "to attIend a less6 expensive school" as a reason for
changing schools: by SES.. .......................... .. 41 23. Percentage of sophomore transfers indicating "to attend a less expensive school"s as a~reason for
changing schools: by SES ............ 41 24. Percentage of students in each type of transferring: by ability (4-year college)............44 25. Group means and standard deviations on academic performance and satisfaction scales .. .. .. 45 26. Standardized regression weights of academic performane and satisfacto with colege on trnsers
:compared to persisters.....................46 27. Percentage distributions of transfers by college cost, type of transfer, and SES .. 47 28. .Percentage of transfer students moving from a high-cost, to a low-cost college by finacl aid and
SES . .. .4.7
A. Background of the Study
Transferring from one college to another has become an increasingly important trend in higher education. A recent national estimate indicated that about 600,000 students move among different types of institutions annually (Willingham, 1973). This estimate includes students who transfer from 2-year to 4-year 'institutions or vice versa, as well as students who move among the same type of institution. This trend of transferring, particularly between 2-year and 4-year institutions, is likely to grow because of the expansion of community colleges and the financial pressures of 4-year college attendance (Watson, 1974; Anderson & Peterson, 1973). In North Carolina, for example, transfers from 2-year to 4-year institutions increased 11.8 percent, and transfers from 4-year to 2-year colleges increased 11.2 percent from fall 1972 to 1973 (Davis & Balfour, 1974). Many other studies have also shown that 2-year colleges have become 'a major source of students for many 4-year institutions (e.g., Willingham, 1972; Trivett, 1974), and that 2-year colleges received as many students from 4-year colleges as they sent (e.g., Illinois Council on Articulation, 1970).
This growingtrend raises several questions concerning transfer students. Of particular interest are the nature .and extent of transferring: who transfers to what type ofinstitution, and for what reason. The information is of value to students, parents, and counselors as well as educational decisionmakers. It may provide a basis for the formulation of admission policies and instructional and financial programs that may help students fuilfill their educational goals. This assistance is particularly important to 2-year college transfers in view of the fact that' more and more students enter 2-year, colleges as they begin their higher education (e.g., Van Alstyne, 1974).
Previous research has provided little information that can be generalized to all institutions of higher edu- cation, since most studies have been limited to a specific institution or geographic region (e.g., Anderson & Riehi, 1971; Hodgson & Dickinson, 1974; Davis & Balfour, 1973). While those studies are valuable to the specific institutions studied, they do not provide a national picture of the transfer phenomenon, nor do they provide a sufficient basis for national policymnaking. A large-scale study involving a representative sample of institutions is a prerequisite to answering questions regarding transfer students in higher education at the natiofiallevel.
In addition, not much is known about the charac teristics of transfer students and how they differ from their nontransferring counterparts (Kintzer, 1973). A comprehensive investigation of the differences between nontransfer and transfer students in background and individual characteristics, as well as their integration into college systems, is needed to provide some information on which national educational programs to meet transfer student needs might be based.
B. Purposes of the Study
The primary purpose of this study is, therefore, to estimate the proportion of transfer students in various categories and to explore some potential explanations of the transfer phenomenon. Specifically, the study was designed to accomplish the following objectives:
(1) To obtain national estimates of the number and proportion of students in various transfer cate- gories;.
(2) To search for variables that could be used to identify students who are likely to transfer; (3) To compare students who transfer from 'a 2-year to a 4-year college with those who enroll in a
4-year college immediately after high school;, (4) To describe students' self-reported reasons for transferring; and (5) To infer from the data some potential explanations for transferring.
1
C. The Data Base.
The data, used in this study were drawn from~the base-year and the first and second. followup data of the National' Longitudinal Study of the High School Class of 1972 (NLS). The NLS data base is comprehensive; its longitudinal design,,based upon a national probability sample, permits analyses thatprvdinomtnabu the psychological, educational, and career deVelopment of people in their early adulthood. The NLS was designed to discover what happens to yIoung people after they leave high school and to relate this information to their prior educational experiences and their personal and biographical characteristics. Educational and work experiences, plans, aspirations, attitudes, and personal background characteristics were measured over three points in time on a sample of over 20,0 hihsho seniors of the class of 1972. The base-year data were collected 'in the spring of 1972, the first followup data were collected in -the fall and winter of 1973-74, and the second followup, data were collected in the fall and winter of 1974-75. Appendix A gives a detailed descrip- tion of the sample, instruments, and data collection procedures.
Of the NLS pricipants who answered the first followup survey, about 50 pecn eeerolled in about 1,800 diverse institutions of higher education in the fall of 1972 (6,196 in 4-year colleges and 3,080 in 2-year colleges). Some of these students failed to provide information about their education in the fall of 1973 or failed to continue their participation in the second followup survey,. and conseqtiently their educational status could not be determined for the fall of 1973 or 1974 and hence their transfer status could not be ascertained. The final number of college students retained for this study was 8,892 (5,974 initially enrolled in 4-year colleges and 2,918 initially enrolled in 2-year colleges). Thus, data about transfer status were available for 96 percent of the students who enrolled in a higher educational institution. There were slightly more men than women, about 52 and 48 percent, respectively. There were about 8 percent blacks, 3 percent Hispanics (i.e., Mexican-American or Chicano, Puerto Rican, and other Latin-American origin), 86 percent white, and 3 percent others. As would be expected of,a college population, the majority of these students were from the famnilies of middle or higher *socioeconomic status (SES)' (only about 1 2 percent of 4-year college students and 16 percent of 2-yiear, college students were from low SES families), from college preparatory high school programs, and had -high academic. ability2 (see table 1).
D. Weighting and Significance Testing
The NLS sample is highly stratified., multi-staged, and clustered. Each case must therefore be weighted by the inverse of its probability of selection to obtain unbiased estimates of Population parameters. Thus, the per- centages, means, standard deviations, and regression 'weights presented in this report are all based upon properly weighted estimates. The standard errors of sample statistics from this complex design are larger than those from a simple random sample~of the same size and should be adjusted accordingly. For example, standard errors of percentages for this complex probability sample can be approximated as a function of the estimated percentage,
SES was based upon a composite of father's education, mother's education, parental income, father's occupation, and a house- hold items index. Factor analysis revealed a common factor with approximately equal loadings for each of the five compo- nents. Missing components were imputed as the mean of the subpop'ulation of which the respondent was a member, defined according to cross.-classification of race, high school program, and aptitude. The available standardized components, both imputed and nonimputed, were averaged to form an SES when at least two nonimputed components were available. The continuous SES score was 'then assigned to one of the quartiles on the basil of the weighted frequency distribution of the composite score. The first quartile, the middle two quartiles, and the fourth quartile were respectively denoted as the low, middle, and high SES. In some analyses, the continuous SES score was used.
2The ability measure was a composite score of four tests: Vocabulary, Reading, Letter Groups, and Mathematics. A factor analysis revealed a general academic ability factor that was represented by an equally weighted linear composite of these four standardized tests. The composite score was classified into a low, middle, or high category corresponding to the first quartile, the middle quartiles, and the fourth quartile.
2
Characteristics.coeg.cleg
SES Low .............. Middle.............41.53 Hfigh........... Unknown .0.0........
High school program Generaf.......19.02 Academic ............ Voc tech............ Unknown............0.02
Ability Low .............. Middle.............27.57 Hfigh...............41.31 Unknown............26.00
Sample N.............
4-Year 2-Year
:21.60 29.70 23.17 28.42 24.83, 13.23 30.40
the sample size, and the estimated design effect, which is the ratio of the sampling variance of the statistic for the sample to the sampling variance of the statistic for a simple random sample of the same size. Thus, the approximate standard error of percentages in this paper can~beobtained by the following formula:
S.E. (P)D (In where p is the percentage, D is the design effect, and n is the actual sample size (see Kish, 1957; Kish & Frankel, 1970). The average design effect for this study is estimated to be approximately 1.35; thus the usual standard errors should be mutiplied by 1.35, which is about 1.16.
To, contrast two subpopulation percentages, d.= pi - p2~,:the standard error of the difference.may be approximatedby taking the square root of the sum of the squares of the standard errors for pi and p2 . The' approximfation will be conservative because of the exclusion of the covariance term for pI and P2 ~in the esti- mation formula.yhi comparing two subclasses of students; the covariance term tends to be positive because of the 'positive correlation caused by the sample clusters of; 18 students per school. The; effect of this positive correlation is to reduce the standard error ofthe difference.
The significance tests of percentages and associated probabilities employed in this report are based on the normal approximation to the binominal'distribution. It should be noted that the ap roximationmano be good for small sample~sizes or extreme percentages.-
E. An OverviewL of the Remnainder'ofthe Report
The remainder of this report is organized according to the objectives described previously. Chapter; II describes the extent of transfer in terms. of percentages and estimated numbers for variou transfer groups. in addition, differences in transfer rates among subgroups, are described (e~g., groups defined by sex, race, and levels of ability and educational aspiration). Chapter III focuses on the differencesbetween transfers and nontransfers in 4-year and 2-year institutions. The comparisons include those between transfIers and per- sisters, and between transfers and withdrawals. Chapter: IV compares vertical transfers (i~e., students who moved from 2-year to 4-year colleges) and 4-year native students on b~ackground variables, financial aid status, satisfaction with college education, and academic performance. Chapter V. follows with tabular summaries of students' self-reported reasons for changing schools. Tabulations are presnesprtlyb type of transfer and type of college. Chapter VI presents tests of several hypotheses related to reasons for transferring; these center on the issue.-of an, incongruency between the student and the institution. The last chapter, Chapter VII, summarizes the major findings and discusses -the implications. Additional information given cursory treatment in the text has been included in the appendixes.
4
II. EXTENT OF COLLEGE TRANSFERS
A simple but significant question about college. transfers is what proportion of students transfer~,and what is their transfer pattern? Of particular interest is the proportion of 2-7year college students who transfer'to 4-year~ institutions. This proportion may 'rev-eal a ~predictable- soflide. of student, enrollment.for the 4-year institutions. Previous studies have not provided a. consistent national picture about college transfers. For example,. one study. (Hoistrom & Bisconti, 1974) found that about 52. percent of~full-tim6 2-year college students transferred to. 4-year institutions over a 4-year period, while another (Van Alstyne et. al., 1973) found that about 36 percent of 2-year- college full-time students transferred to 4-year colleges over, a simiar:tm eid ut(92 ni ,cated that new transfer ~students in 1968 numbered about 456,000, while Willinghanm (1972) estimated ~the. number to 600,000 annually. The inconsistencies may reflect,the change of college-goingtrends in recent years, .or they may reflect the nonrepresentative samples of institutions. To meet this need, 'two questions are addressed in this chapter: What percent of American college, students, move among institutions4of higher education annually? Are there differences iii transfer rates among subgrous defined by institutional charac- teristics and by personal backgroundvariables?.
To answer these 'questions,- various :categories fo' college transfer students were defined. Based upon'-edu& cational~status in October 1972, 1973, and 1974, students were classified into 'persisters, transfers,:'and with-' drawals. Detailed tree diagrams, including the percentage of students at each decision point for those students" enrolled in a 4-year college or a 2-year college, are presented in appendix B.
The transfer students were further; divided into the following categories: (1) 4->2 Transfers: students transferring from a 4-year college~to a 2-year college, often called ~reverse
transfers in the literature; (2) 2->4 Transfers:- students transferring'from ~a 2-year college to a 4-year college, often~called vertical~
transfers; (3) '4->4 Transfers:~students transferring from a 4-year college to aohr.4-yea college; and6 (4) 2->2 Transfers: students transferring from a 2-year college to another 2-year cIollege.~These last
two categories are often called horizontal transfers. The numerical labels were used to designate transfer categories 'for clarity and to avoid the Value-judgment connotations implicit in such terms as reverse and vertical.,
A. Total Transfer Rates
1. Transfers in the First Year
Many students moved among colleges during or at the end of their first year of matriculation. The pe'rcen- tage of transfers, based upon, initial total enrollment in 4-year or 2-year colleges, is show Iin figure 1. About 8 percent of 4-year college students moved' to other 4-year institutions,. and about 3 percent moved to 2-year colleges. During the same period of time, about 6 percent of 2-year, college students moved to 4-year colleges, and about 3 percent moved to other 2-year colleges. Itis estimated that a total of 142,141 (the total sum of the four transfer categories), of the high school seniors of 1972 who enrolled in~collIeges by October 1972 transferred by October 1973. This indicates that 1 out of 1 0 studentsimoved during the first year of colIlege. The 4->4 transfer group was the largest,: and the 2->2 transfer group was the'smallest, in terms of both percen- tage and actual number oftransfers.
An interesting point should be noted; that is, the number of'4-*2 transfer students was about the same as the number 'of 2-4 transfer students (see 'Figure 1). Thissupports previous finidings that the 2-year colleges receive as many students from the 4-year colleges as they send (Illinois Couni nAtclto,17)
5
73,313
28,073 2,6
college-N coe%,
14,587 (3.31%/)
Figure 1. Percentages and estimated national totals of students who transferred during or at the end of their first year in college. (Initial college is represented by shaded circle.)
NOTE.-- Sample N, for 4-year bollege initial enrollment was 5974, and for 2-year college it was 2918.
2. Transfers in the Second Year
Many students remained in the same college for more than one year and then transferred to another college. As would be expected, this was especiall common among 2-year college students. Based upon the initial enroll- ment of October' 1972, about 17 percent of 2-year college students transferred to 4-year institutions during or at the end of. their second year (see figure 2). The percentage of 2-*4 transfers based upon, sophomore enroll- ment was greater (about 29 percent). In either case, a greater number of 2-year college students transferred to 4-year colleges during or at the end of the second year than during the first year.
Transferring between 4-year colleges was still substantial during, or at the: end ofthe sophomore year. The pretage .was about 6 percent based on the initial enrollment, and about 9 percent based on the enrollment of
sophmornotrasfe stdens. This indicates that proportionally there were as many 4-*4 transfers in the secodyars i th fist earof college.
The4->2trasfer mad upabout 1 percent, based,upon the initial enrollment. Although small, thisgru of tudntsis artculrlyintrestingbecause they could be expected to have completed a 2-year college degree
by this point in time if they had started at a 2-year institution. Their -reasons for transferring,ae discussed in chapters V and VI.
dverallt it is estimated that a total of 146,770 (the total of the four. transfer categories) of the high school seniors of 1972 who enrolled in colleges by Qctober 1972 transferred during or at the end of the sophomore' year. The~2-4 transfer group was the largest, and the 2->2 transfer group was the smallest in terms of actual number of transfers (See figure 2)..
6
[1.3211,1 t~~~~~~~~[29.34%]
4,011 (0.9 1%7) [1.54%]
Figure 2. Percentage and estimated national totals of students who transferred during or at the end of their second year in college. (Initial college is represented by shaded circle.)
NOTE.-- based upon the initial enrollment. []=based upon the enrollment of those who persisted for one year.
Sample N for 4-year college initial enrollment was 5974, and for 2-year college it was 2918.
3. Total Transfers Over 2-Years
The estimation of the total percentage and the number of students who transferred among colleges over a 2-year period requires further consideration of the changes of student college-going status. The four transfer' groups are further specified in table 2. The specifications indicate the type of initial and final colleges.-Thus, a student who entered a 4-year college, transferred to a 2-year college, and then transferred back to a 4-year college whould be indicated by a 4-*4 transfer, as would a single transfer between 4-year colleges. If other classification schemes are of interest, the estimates of percentages -and 'numbers can~be obtained from appendix B.
The total percentages and numbers of students for the transfer groups are, summarized in figure 3. Two years after initial matriculation, slightly over 24 percent, of 2-year college students transferred to 4-year colleges. (It ,should be noted that about 52 percent left school, and 24 percent were still in 2-year colleges.),Those trahsfers constituted about 14 percent of theItotal 4-year college enrollmient. (This was calculated on the basis of estimates presented in Appendix B.) The proportion of 2-year college students who transferred to 4-year institutions was consistent with findings of some previous studies (e.g., Van Alstyne, 1974). However, the total number of transfers was smaller than that estimatedby Willingham (1972), based upon regional or institutional studies.
Table 2.-Transfer specification
1. 4-4 transfers: 1. 4-+4 ~~ transfers:~~ 4 -
4 -
~~O~
2
2
2
2
2
IW = Withdrawing from study
8
About 3 percent of the. 4-year college students moved to 2-year colleges. This figure did not differ much from the first year's 4-*2 transfer rate because about a quarter of the first-year group went back to 4-year colleges, and about a quarter withdrew from colleges (see appendix B). Figure 3 also shows that about 16 per- cent of 4-year college students moved among 4-year colleges over 2 years. The total number ofsuch 4-*4 trans- fers was slightly greater than that of the 2-*4 transfers.
143,261 (1 6.09%)
17,728 (4.03%)
Figure 3. Total percentages, and estimated number of students who transferred over 2 years. (Initial college is indicated by shaded circle.)
NOTE.-- Sample N for 4-year college students was 5874, and for 2-year college students it was 2918.
B. Transfer Rates by Subgroups Defined by Background Variables
A question Of interest is whether .there are differenIces in transfer rates among subgroups defined by back- ground variables. This section presents transfer rates for varying subgroups and describes their differences. However, the primary focus of this section is to describe group differences. hin later chapters, selected variables will be consideredjointly in more detail.
The following background variables are included in the analyses: sex, race, socioeconomic status (SES), aptitude, educational aspiration, high school prgagoraphical region of high school where graduated, college grades, and field of study. Geographic region was used as a variable because there were more 2-year colleges available to residents of the Western region, and relatively more students 'in the South and West than in the Northeast or North Central were enrolled in 2-year colleges (see table 1). Consequently, it would be more likely for students in those two regions than in other regions to transfer from 2-year to 4-year institutions or vice versa. Other variables~were selected because they reflect students' background characteristics (sex, race, SES), individual attributes (aptitude, aspiration), high school training (high school program), and college experience (college grades, field of study)-variables that might be related to college transfer behavior.
9
The transfer rates presented i n the following descriptions are ~the total transfer rates, over 2 years. This choice is particularly approIpriate for 2-year college~students because, to many of these students, the second year is the final year, and transferring is necessary to continue higher education. As previously defined in table 2, the transfer designation indicates the type of the initial and final colleges. Thus, a student who entered a 4-year college, transferred to a,2-year college, and th-en'transferred back to a 4-year college would be 'indicated by a 4-4 transfer, as would a single transfer between 4-year colleges.
The percentages of students who transferred by October 1974 are summarized in tables 3-a and 4-a, respec- tively, for the 2-year and 4-year institutions for subgroups formedby nine variables. ~The tests of significance for subgroup differences are p resented in tables 3-b and 4-b. Several patterns of group differences can be seen:
(1) There were no substantial differences in any of the four transfer rates between mhen and women. This finding does not support earlier findings that men are more likely than women to transfer, particularly from 2-year to 4-year colleges (e.g., Holmstrom & Bisconti, 1974). Th inconsistency could be due to the lack of representative samples in the previous studies or to a different time period (e.g., 4-year time span in Holmstrom and Biscontli's study), in which more men than women reentered colleges after a few years of work. Nevertheless, the current finding of no sex differences in the 2-~4 transfers may indicate that more women than before are becoming career-oriented and desire higher education.
(2) Differences in the 4->4 transfer rates among several subgroups were significant. As shown in table 3-b, whites were more likely than blacks to transfer; ,students of high SES were more likely than students of low SES to transfer. Likewise, students of higher educational aspiration and higher college grade-point average were more likely to transfer~than those with low aspirations and averages. In summary, the groups more likely to transfer are characterized as being white, of high SES, academic high school program participants, high aspiration, and high college achievement.
(3) Differences in the 4-~2 transfer rates existed between the West and NoIrth Central regions. This is probably because there are more 2-year colleges in the West than in the North Central region, and thus there are more opportunities for students in 'the West to move from 4-year to 2-year colleges. Another difference in the 4-*2 transfers existed between low and' high achievement groups; Stu- dents having lower college grade-point averages were more likely to transfer from 4-year to 2-year colleges. This finding lends support to an argument that many 4-year college students-intend to improve their grade-point averages in a 2-year college, and then contminue in a 4-year college (Kuznik, Maxey, & Anderson, 1974).
(4) There were no substantial group differences in the 2->2 transfer rates; that is, students of this sort did not concentrate in any subgroup defined by the selected background variables.
(5) Differences in the 2->4 transfer rates were evident amon.g all subgroups except those defined by sex. As shown in table 4-b, whites had a greater 2-*4 transfer rate tha~n blacks, and blacks had a greater transfer rate than Hispanics. The South had the highest and the West had the lowest 2-~4 transfer rates. The reason why the West had the lowest 2->4 transfer rate, as opposed to the highest 4-*2 transfer rate, is unknown. It may be. due to a greater proportion of Hispanics living in the West than other regions;, Hispanics had the lowest 2->4,transfer.rate and the highest 4-÷2 transfer rate among the race groups. Other .group differences were, in an expected ,pattern. Students in academic fields and: students of higher SES, ability, aspiration, and college academic performance had a greater 2-~4 transfer rate than students of lower levels on these variables.
10
*Table 3-a. -Percentage of 4-year college students who transferred by the end of- the sophomore year:, by subgroups
Subgroup j 4.-+4 ~~~Transfers
4-+2 Transfers
14.13 16.96
.. .. ....
.. ... ... ..
.. .. .... ...
College grade > A-...... .. ... B+ to B-...... ..... C+ to C-.........13.98 <C-...........9.19
20.57 18.47
I1I
Table 3-b.--G~roup differences in transfer rates (In percent) from 4-year colleges
Sex: Male-female.............-1.92 0.44
1.55 3.44
6.42t -6.09t
.6
-2.34 .3.60*
-.87 -1.26
Educational aspiration: < coil, to > 4 yr. coil..... 2yr. coll.to >4yr..... ...... < coil. to 2 yr. coil..........
4i......1.89* 11.39*
Nonacademic-academic.......4.34.3
College grade: (>A-) to (B+ to B-) .......... (>A-) to (C+ to C-) .......... 16.59* (->A-) to (<C-)............11.38* (B+ to B-) to (C+ to C-).........4.49* (B+ to B-) to (<C-) .......... (C+ to C-) to (<C-) ...... a....4.79
2.10
9.28*
.- 1.18 .3.58* .5.67 * -2.40 4.449* 2.09
* p<.Oi (a two-tailed test). t not significant at the .01 level because of greater standard error.
12
Table 4-a.-Percentage of 2-year college students who transferred by the end of the sophomore year: by subgroups
Subgroup 2+ Transfers
6.80 3.90
9.08 26.05
84.12 70.04
179 2,279
Middle........ High.
13.84 5..05
22.78 31.95
73.38 63.01
1,539 789
4.56: 3.95~
4-year college...... 4.92 4.26
College grade >A-. ......... B+ to B ......... C+ to C .........
<C-..........
13
Table 4-b. --Group differences in transfer rates (in percent) from 2-year colleges
Group comparison
Sex: Male-female.............
Ability: Low-high............... Middle-high ............. Low-middle ............
Region: Southwest.............- North central-west...........-. North~east-south ............ North central-south..........
Educational aspiration: <Coil,. to >4 hr. coill..........-1.88 2-yr. Coll. to >4 yr............66 < coil. to 2-yr. coil...........-
Field of study: Nonacademic-academic.........-.59
College grade: (,>A-) to (B+ to B-)...........-.04 (zA-) to (C+ to C-) ...... I....-2.20 (~_A-) to (<C-)............-2.77 (B+ to B-) to (C+ to C-).......I...-2.16 (B+ to B-) to (<C-) .........- (C+ to C-) to (<C-)-..........-.57
Difference in transfer rates
14
C. Transfer Rates by the Type of Institution
This section presents percentages of transfers by type of institution. Institutions may be characterized by length of Program in years, type of control (e.g., public versus private), size, and selectivity level.
1. LengthofProgram inYears
As shown previously, transfers were defined separately for 4-year and 2-year institutions because of dif- ferences in the nature of their curricula. According to the count at the end of the second year of enrollment, there was a greater proportion of 4-4 transfers than of 2-*2 transfers (about 16 percent versus 4 percent). The majority of the transfer students from the 2-year colleges were movingto the 4-year colleges (about 24 percent of the initial total enrollment). On the other hand, only about 3 percent of the 4-year college students trans- ferred to the 2-year colleges (see figure 3).
2. Type of Control
Several studies have shown that a larger proportion of students from private than from public colleges trans- ferred to other institutionswithin a 4-year period (e.g., 'Hohmstrom & Bisconti, 1974; Van Alstyne et at., 1973). The NLS data supported this finding. As shown 'in table 5-a, the overall transfer rates were significantly higher for students from private institutions. (T'hese rates were based on those individuals who entered college by October 1972 and who transferred sometime during the ensuing 2 years.) Specifically, about 19 percent from the 4-year private institutions transferred to other 4-year schools, compared to about 15 percent of public college students. Students from private 2-year colleges had a 2->4 transfer rate of about 35 percent, compared to 24 percent of students from pulc2ya ntitutions. Both 4-~2 transfers and 2->2 transfers were in the same direction-private institutions having a greater percentage than public institutions;however, the differences were not significant.
Table S-a.--Transfer rate(in percent) by type of institutional control
Controlof
4->2 I Non- 2->2 2-year college
2-4 I Non- I Transfers Transfers transfers N Transfers Transfers transfers N
Public. 14.79 3.09 82.12 4,004 3.75 24.12 72.13 2,575 Private.... 19.18** 3.27 77.55 1,597 8.35 34.52** 57.13 173
**Students at private institutions had a significantly greater transfer rate than. those at public institutions (pc.'01, a one-tailed test).
A related, question is what percentage of students transferred from a public to, a private institution, and vice versa. To answer this question, students who transferred during or at the end of their first year of matriculation were cross-classified by the type of control of their initial and destination colleges. Results indicate that the majority of private as well as public college transfer students moved to public institutions. for example, about 61 percent of 4-*4 transfers and 92 percent of 4-*2 transfers from private institutions moved to public institu- tions whereas only about 26 percent of 4->4 transfers and 3 percent of 4-*2 transfers from public institutions moved to private institutions. A similar pattern existed among transfer students from 2-year colleges (see table 5-b). This phenomenon seems to 'indicate that financial and/or academic pressure could be an important factor in the transferring process since private institutions are more competitive and expensive than, public institutions.
15
Table 5-bx--Transfer students cross-classified by type of control of initialand destination colleges
Percent control of Transfer category
Control of initial collegeN
Public . 96.92 30 3
Private....... . 91.55 8.45 40
2-4 Public....
Private........ I.....79.17
53.01 ~ 20.83
46.99, 148 25
3. Selectivity Level
Analysis of the data by selectivity and size of the institution is another approach to describing transfer rates. Information about the institution's selectivity level and size was obtained in part from sources other than the NIS data. A preliminary analysis, using Astin's (1971) college selectivity index' with eight levels and college size (Suchar, Van Dusen, & Jacobson, 1974) with five levels, is discussed below. The sample size was reduced, since not all colleges had the supplementary information.
Transfer rates did not var in a linear manner with the selectivity levels of the 4-year institution. In colleges of selectivity levels 1 through 6, the 4->4 transfer rates were generally in an ascending order (table 6-a). How- ever, students from institutions of selectivity levels 2 and 7 had lower percentages of transferring than students of other institutional levels. As for 4-+2 transfers, there were almost no differences, except that students from the highest selectivity level had the lowest percent age of 4-*2 transfers. It is noteworthy that less than 1 percent of students in 4-year institutions of highest selectivity (level 7) moved to.2-year institutions.
Transfers from the 2-year institutions were not included in the selectivity analysis because only a few such institutions had a selectivity level greater than 3; consequently, little variability would be expected across so few levels.
It was concluded that the 4-year college-transfer rates were not linearly' related to the institutional selec- tivity level; the transfer rates of students from the more selective institutions were not necessa rily higher than those from less selective institutions.
Another aspect of the transfer pattemn relating to selectivity level is the proportion of students who trans- ferred from low to-high selectivity institutions, and vice versa. Based upon available data, .about 64 percent of 4-year college transfers from high selectivity colleges (levels 4 to 7) in October 1972 moved to, low selectivity colleges (below level 4)~in October 1973, and about 36 percent moved to colleges of simi'lar selectivity levels. Of those transfers whose initial colleges were of low selectivity, 23 percent moved to institutions of high slc- tivity, and 77 percent moved to colleges of similar selectivity (see table 6-b). The higher proportion,of students moving from high to low selectivity colleges may indicate that competitiveness is a factor in the transferring process.
Selectivity index is based upon the average SAT and/or ACT scores of the, entering students. There are eight levels,of selec- tivity, 1 being the lowest and 7 being the highest level, and 0 (unknown) indicating that no direct estimate of selectivitywas available. In general, the "unknowns" tend to be around levels 1 and 2 (Astin, 197 1, p. 24).
16
Table 6-a.-Transfer rates (in percent) of 4-year colleges by selectivity level of institutions
1
Selectivity 4I44->2 Non-Nlevel SelecivitylevelTransfers Transfers transfers N
Unknown 0. .. ..... 17.54 3.41 79.05 221 LOW 1........16.24 2.99 80.77 461
2.........12.71 2.79 84.50 586 3........16.61 3.83 79.56 826 4....... 16.68 3.20 80.12 952 5....... 18.56 2.07 79.37 546 6........18.11 3.23 78.66 213
High 7........ 11.73 .77 87.50 205
NOTE.--Nontransfer includes persisters, graduates, and withdrawals.
Table 6-b.--Initial and destination college selectivity level of 4-year college transfer students
Selectivity of Selectivity of destination college initial college in October 1973 (inlpercent) N inOctober 1972
Hfigh Low
Hfigh .............. 35.86 64.14 198 Low .............. 23.14 76.86 363
Note.--(1) High - Selectivity levels are greater than or equal to levelA4 Low - Selectivity levels are lower than level 4 or are unknown.
(2) 4->4 and 4->2 transfers were combined because of small sample size and small number of high selec- tive 2-year colleges.
4. Size of Institution
The size of insitutions seems to be related to transferring. As shown in table 7-a, students from the larger 4-year institutions had lower percentages of transfers than those from smaller institutions. This suggests that large institutions have greater holding power than smaller ones, probably because of greater variety of programs and social opportunities. The differences in 4-*2 transfers were not in a linear pattern, however. Students from institutions over 15,000 had the highest 4#+2 transfer rate.
Unlike students in the 4-year institutions, students fromlre2ya instttoshdahge -2tase rate than students from smaller,colleges. This trend, however, was not shown in 2->4 transfers;- both small and large institutionshad a higher vertical transfer rate than institutions of middle size (see table 7-a).
17
Table 7-a --Transfer rate (in percent) by size of institution
<2,000 *19.89 3.14 76.97 1,154 3.58 32.44 63.98 853 2,001 - 5,000 16.90 2.40 80.70 1,011 4.09 19.24 76.67 646 5,001 - 10,000 14.09 3.41 82.50 *935 5.44 16.44 78.12 253
10,001 -15,000 13.08 1.71 85.21 397 5.01 18.26 76.73 145 > 15,000; 12.51 4.19 .83.30 497 7.13 24.76. 68.11 88
Transfer students were cross-classified by the size of their initial and destination colleges. The classification did not reveal any consistent transfer pattern. Students were not necessarily moving from large to small colleges or vice versa. Although the majority of 4-+2 transfers moved to small colleges, this may simply indicate that 2-year colleges are generally small (see table 7-b).
Table. 7-b.--Transfer students cross-classified by size of initial and destination colleges
Transfer Size ofinitial category college (1972) <200 2,001- 5,001- 10,001- N
4-~4 <2,000 33.26% 13.29% 26.35%o 14.24% 12.85% 83 2,001 - 5,000 19.16 18.08 31.43 16.48 14.84 63 5,001 -10,000 25.72 17.11 33.60 17.08 6.49 53
10,001 - 15,000 40.82 9.25 8.72 24.43 17.79 1 8 >15,000 13.07 22.63 31.50 5.68 27.11 17
4-2 ~ <2,000 55.16 28.32 13.54 0.00 2.97 2 2,001 -5,000 48.34 41.84 0.00 9.82 0.00 13 5,001 -10,000 51.27 35.93 12.80 0.00 0.00 15
10,001-15,000 ----- --
> 15,000 ------ -
2-+2 <2,000 57.92 17.97 0.00 5.97 18.13 20 -5,0002,001 22.14 43.79 25.03 9.05 0.00 16
5,001 -10,000 0.00 28.56 66.97 0.00 4.47 1 3 100115,000 --- -- -
*>15,000 - -- --
2-4 ~ <2,000 16.40 14.20 22.91 25.36 21.13 52 2,001 - 5,000 22.60 17.71 16.78 26.19 16.72 22
* 5,001 -10,000 - --
10,001 -15,000 - --
>15,000 -- --- -
NOTE.--Symbol -- indicates that the N's were too small for reliable estimates..
18
D. Summary and Discussion
The extent of college transfers was investigated by estimating the national proportion of college students in four transfer categories: (1) 4-*4 transfers, (2) 4->2 transfers,: (3) 2->2 transfers, and (4) 2->4 transfers. ~Differences among subgroups defined by background variables aid 'institutional characteristics were also included.
The number of transfers from 4-year colleges was substantial. About 19 percent of4-year college students transferred within 2 years after initial matriculation (see figure 3). Of those. transfers, 84 percent moved to other 4-year colleges, and 16 percent transferred to 2-year colleges. Proportionally, there were as many 4-*4 transfers in the first year as 'in the second year. Of the 4-~2 transfers in the first, many might,move back to a 4-year institution in the followingyear (see figures 1 and 2).
Transfers from the 2-year colleges were also substantial. About a quarter of the students transferred to a 4-year institution over a 2-year period. A majority of those students did so in their second year (see figures 2 and 3). The number of 2->2 transfers was least substantial among the four transfer groups.
The observed transfer rates were; 'in general, smaller than those found by other studies. It is possible that these estimates are smaller because ,manymore students may transfer to or reenter colleges in subsequent years. A more accurate, estimate of 2-*4 transfers, for example, requires data covering a longer time span. The next NLS~followup will be valuable in this respect.
The 4-+2 transfers were somewhatunconventional. Although some of those students may eventually return to 4-year colleges, the large number of thos students, as shown by the NLS data and data from Illinois (Illinois Council on Articulation, 1970) and North, Carolina. (Davis & Balfour, 1973), point to the need for counseling services in college selection, and perhaps in curriculum programs. On the other hand, the phenomenon also suggests that 2-year colleges play an important role, in higher education. They are a mobility channel for the lower SES student, the late bloomer academically, and the less college-degree-aspired student. In addition, many students may redirect theft goals, as well as improve their academic standing, in 2-year colleges.
Differences among subgroups existed primarily in the 4->4 transfers and the 2-*4 transfers (see tables 3-b and 4-b). Students of high SES or high ability were more likely to move from one 4-yearcclege to another. In a similar manner, students of higher SES, aspiration, achievement, and/or ability had a greater 2->4 transfer rate than students of lower levels on these variables. A further investigation of the relationship between background variables and transfer rates is discussed in the followingchapter.
To a large extent, the findings on the differences in transfer rates among the types of institutions were con- sistent with previous studies. Students from private institutions had a greater transfer rate than did students from public institutions.The difference in transfer rates between public and private institutionIs may be partially due to the selection of different kinds of students, as well as to different institutionalenvironments. Four-year private institutions tend to be selective., and the resulting competitive pressure may lead some more motivated students to transfer to other institutions. Private institutions also tend to be more expensive. On the other hand, many public institutions, are large State-controlled schools which are able to provide a greater variety of sub- cultures for students to identify with. As Kamnens (197 1) argued, larger institutions exert greater holdingpower over students by providing more diverse programs and social activities; a greater variety of opportunities leads students to greater commitment.to the 'institution, which, in turn, results in less transferring from the insti- tution. The NLS data only partially support the above arguments.
The relationship between college selectivity levels and transfer rates was not significant; transfer rates did not vary in a consistent manner across selectivity levels (see table 6). It is possible that the feeling of prestige in a highly selective institution may counterbalance the pressure of competition. However, when students in highly selective colleges transferred, a great proportion of them transferred to less selective institutions.
The differences among -institutions of varying sizes, however, ~showed a consistent pattern; the larger the school, the smaller the 4->4 transfer rate. A larger school seemed to exert a greater holding power over students.
III. COMPARISONS BETWEEN TRANSFER AND NONTRANSFER STUDENTS
A question posed in this study is what are the characteristics of transfer students? In particular, are trans- fers different from nontransfers in certain background variables? Other studies have asked this same question (e.g., Slettedahl, 1972; Willingham & Findikyan, 1969; Van Alstyne, 1974; George et al., 1973); and it has been found, for example, that men are more likely than women to transfer from 2-year to 4-year institutions-(Van Aistyne, 1974), students of high SES and high aspirations are more likely than students of low measures in these variables to transfer (Kintzer, 1973; Brinbaumi, 1970), and majority students are more likely than minority stu- dents to transfer (Willingham, 1972). To a great, extent, descriptive analysis in the preceding chapter has pro- vided some, supportive evidence. However, since many background characteristics are intercorrelated, an observed simple relationship between predictor and transfer behavior may disappear when other variables are controlled. In addition, studies did not include Withdrawals and graduates (i.e., students who completed the 2-year program but discontinued further study) as comparison groups, and thus much information may have been lost. It is, therefore, the primary purpose of this chapter to further examine the differences in student characteristics between transfer and a more refined nontransfer group* as well as the relationship of a back- ground variable with transfer behavior when other variables are considered.
The college-going status of the students who initially enrolled in 4-year or 2-year institutions by October 1972 was examined again in October 1973 and 1974.- This examination provided a basis for classifying the students into the following categories: ~persister, transfer, graduate, and withdrawal. Transfer groups have been specified in the previous chapter (see Table 2). Persisters are those students who remained in the same college from October 1972 to October 1974. Withdrawals were those students who were in school in October 1972 but were out of school by October 1974. Graduates were those 2-year college students who hdcompltd a 2-year degree but did not continue their education in October 1974. The student categories are further listed below:
Four-year institutions Two-year institutions
Transfer .J~~~i.4-*4 transfer 1.2-~2 transfer f2. 4-~2 transfer 2.2-~4 transfer
J3. Persister 3. Persister Nontransfer 1.4. Withdrawal 4. Withdrawal
I...Graduate The comparisons between transfer and nontransfer;students were made on the following background vari-
ables: socioeconomic status, sex, race, high school grades, aptitude test scores, educational aspiration, high school program, college grades, hield of study, and region. These variables were also described in the previous chapters. I
The primary purpose of this analysis was to compare a transfer group with a specific nontransfer group, rather than to test the overall differences ambong student groups. Thus, the analyses were the so-called planned comparisons on the selected groups, and the same error term (within-group varianewausdfra1tt. The comparisons selected for the 4-year and 2-year college students are listed beow. It should be noted that* the number of comparisons allowed for each set of analyses should not be greater than K-i, where K is the number of groups.
For the 4-year college students, the comparisons were: 1. Persisters vs. 4->4 transfers 2. Persisters + 4-*4 transfers vs. 4-~2 transfers 3. 4--2 transfers vs. withdrawals
For the 2-year college students, the comparisonswere: 1. Persisters vs. 2->2 transfers 2. Persisters + 2-~2 transfers vs. 2e4 transfers 3. 2-~4 transfers vs. graduates 4. 2-*4 transfers vs. withdrawals
21
Multivariate analyses of variance were performed separately.for the 4-year and 2-year college students on 12 variables. The first step involved the computing of the weighted means and variance-covariance matrix, which were then used together with the actual sample n's as input data for analysis. (The requirements ofthe weighting process were described in chapter 1, section D.) ,Four sets of test statistics are .presented for each comparison: the multivariate F-ratio for the overall group differences. on the variables simultaneously; the uni- variate F-ratio for the significance of the individual variable; the step-down..F-ratio for the test of an individual variable by holding prior variables constant; and ~discriminant functions for providing the maximum differen- tiation between groups. The standardized discriminant function coefficients have an interpretation analogous to that of beta weights in a regression analysis;,that is, they not only indicate the relative partial contribution of a variable holding other variables constant, they also indicate the direction of the effect. It should be noted that, because of the unequal sizes of student groups (i~e., nonorthogonal design), each comparison of interest was placed in the last position to obtain unconfounded tests (sdte Bock, 1975; Finn, 1974).
A. Comparisons Between 4-Year College Transfers and Nontransfers.
The weighlted means and common standard deviations (i.e., pooled across groups) of the background vari- ables are presented in table 8. Several variables were zero-one dichotomies for which the means are proportions of students having the, related background characteristic. For example, the value of .47 in the first column of table 8 indicates that 47 percent Of persisters were female students. All continuous variables such as SES and aptitude test scores were coded from low to high. The test statistics for the three comparisons are presented in tables 9-a, 9-b, and 9-c, and-are discussed below..
Table 8.- Weighted means and standard devitat ions for various college-going status groups on background variables (4-year colleges)
Background variables' Persisters. Withdrawals sadrTransfers Transfers dvain
SES...............2A .41 .24 .69 Female (vs. male) .. ... I.....47 .57 .54 .49.5 Black (vs. nonblack) .. ..... I.... .06 .04 .05 .06 .23 1Hispanic (vs. non-Hfispanic)...... .01 .01 .04, .01.1 High school grades.. ......... 6.65 6.65 6.30 6.13 1.15 Academic aptitude test scores....58.26 57.85 57.49 55.79 5.73 Educational aspiration .. .. .... 5.68 5.64 5.6.1 5.39 ..62 Academidc high school program (vs. general
and voctech) ............ .84 .82 .82 .69 .39 College grades............ 5.79 5.93 5.04 5.09 1.32 Academic field of study (vs. nonacademic). .95 .95.- .92 .87 .25 South (vs. others) ........ ... .25 .24 .33 .29 .44 West (vs. others) .. 14.5 .28 .16 .35
N'.......... 1948 547 98 653 3246
SES is a composite score with a mean of 0 and standard deviation of 1. Aptitude test scores are standardized scores with mean of 50 and standard deviation of 10. High school and college grades were coded as fonlows: mostlyA -_ 8; about half A and half B:= 7; mostly B = 6; about half B and half C = 5;mostly C =4; about half C and half D = 3;-mostly D = 2; and mostly below D = 1. Education aspirations were coded as follows: less than high school =_ 1; high school =2; some vocational studies beyond high school= 3; two-year college =4; four-year college = 5; and graduate school = 6.
'The squares of these values are within-group means of squares (the error terms for univariate analysis). 3 The differences in sample size 'in this analysis and previous analyses were due to missing data on backgroundvariables,
primarily because of nonparticipation in the base-year survey.
1. 4->4 transfers and persisters were different with respect to their overall background (the multiyariate F-ratio of 2.83 was significant at the :001 level with 12 and 3231 degrees of freedom, see table,9-aix The dif- ferences were particularly substantial in SES, sex, -and college grades ~(see* the ~univariate EF-ratios for ~these variables in table 9-a). The differences on these variables still existed even when some priorvariables were con- trolled (i.e.,~the stepdown F-ratios on thes~ variables were-still significantat the .05 level).. After SES, sexj,race, and high school grades were considered, persisters had significan~tlylhighertest scores than transfers..,.-
The discriminant function coefficients show that the variables of SESsex; aptitude test, and college grades: carried greater weights than other variables in1 differentiating-thepersisters from 4-t4 transfers;-As indicated by: the -sign of the coefficients and statistics in table 8, the. 4->4 transfer. group: was composed 'of more female students than the persister group. This indicates that more female tan male students transferred among 4-year.- colleges, or male students were more likely than female students to remain in the same 4-year college. (Note:. The proportion of female students in the initial 4-year college ,enrollment was .48;~ see table ,1.) The 4->4 transfer students also tended to have higher scores on SES and college achievement than persisters after other variables were considered. The groups were about. one-tenth dff a standard deviation apart on both variables. However, it should be noted that 4->4 transfer students had lower aptitude test scores than persisters. It may be possible that the lower high school grades and- aptitude test scores of thos:e 4-44 'transfer students prohibit them from getting into' the kind of institution t hey like, and transferring becomes an alt~ernativesolutioni.
Table 9-a--Test statistics for the comparison between' pMiters ad4->4 transfers
Univa Irite F'Stpdown F2Standardized
Variable UiaitF' ted nF 2 discriminant (d~~f. =' 1,3242) coefficients'
SES ..... ..... 406 *4 4.06* 0.44 Female (vs6 male) .. ....... 15.46** 16.29** .66 Black (vs. nonblack). 1.18 .73 -2 Hispanic (vs. non-H-ispanic).......... .00 .09 -.00 High school grades...........00 1.05 -.15 Academic aptitude test scores ........ 2.17 5.02* -.51 Educational aspiration. 1.88 .70............ -14 Academic high school program
(vs. general and voctech)............ 43 .02 -0 College grades...........5.09* .47.5.78*
Academic field of study (vs. academic) . .12 .03 '.03 South (vs. others)............... 22 .09 -.04~ West (vs. others)...........29 .04'.05'
Multivariate F =2.83 X2 (1-2) = 33.81. (d.f. = 12, 3231) p<.0
p<.o 0 1
NOTE.--l1. Within-group variance is shown in table 8. 2. Variables are listed in the order in which, the stepdown analysis was perfonned. Thus,~the stepdown
F shows the significance of the indicated dependent variable, controlling for all variables listed above it.
3. The sig~i of the dliscriminant function coefficients shows the 'direction'of relationship. A positive sign indicates that transfers were higher on the dependent variables than were persisters.
4. *p .c.05; **p<.Ol
23
2., Students who moved to 2'year colleges (i.e., 4-+2 transfers) were different fromn those who remained in t4.year colleges (including 4->4 transfers) in their overall backgrounds. The multivariate F-ratiovwas significant (see table 9-b). The differences were particularly substantial in' the variables of Hispanic versus non-Hispanic, high scho~ol'grades, college grades, and West versus non-West. (The univariate F-ratios for these variables were significant at the .01I or .05 level with 1 and 3242 degrees of free~dom.). Table 8 shows that the 4-*2 transfers were composed of relatively more Hispanics and more students from the West than were persisters, and had substantially lower high school and college grades. The stepdown tests provided the same conclusion for these variables when some prior variables were controlled. In 'fact, as shown by the-sign of the discriminant function Coefficient (see table 9-b), the direction of. lower grades and- greater composition of Hispanic students and~stu- dents from the West still held when' all other variables were considered. In addition', college grades: carried the largest weight in differentiating 14->2 transfer students from those who remainedi-in a 4-year college. It seems that a poor grade-point average was a Major factor leading those students to transfer to 2-year colleges.
___________________________________________
*(~.1, 3242) functionI ~~~coefficients3
SES............... 1.14 1.14 .-. 15 Female (vs. male)...........09 .06 .19 Black (vs. noniblack).. ..... 00 .06 .- '.08 Hispanic (vs. hon-H-ispanic).......4 .58*4 3.96*.1 High school grades.... 9.03** -.26 .9.80**
Academic aptitude test scores ....... 89 .84 .30 Education aspiration...77..177 -.06 Academic high school program
(vs. general and voctech) ... .... 07 .06 .15 College grades ........... 35.27""* 28.60** -.73 Academic field of study
(vs. nonacademic)..........1.28 1.36 -.14 South (vs. others) .. ........ 3.39 3.09 .. 38 West (vs. others) ..... 3.51** 21.86** .59
Multivariate F =5.97 08x
(d~f. =12,3231) .X( 2 =0S p < .Ol . <. 0
NOTE.--l. Within-group variance is shown in table 8. * 2. Variables are* listed in the order in which the stepdow anlsis was performed. Thus, the stepdown
F shows the significance of the indicated dependent variable, controlling for all variables listed *above it.
3. The sign of the discriminant function coefficients shows the direction of relationship. A positive sign indicates that transfers were higher on the dependent variables than were students who re- mained in 4-year colleges.
4. *P <.05; **p< .Ol
3. The third comparison focused on the differences between~those who withdrew and. those who trans- ferred to 2-year colleges (i~e., 4-~2 transfers).. As shown in tables,8 and 9-c, these two groups. of students were distinctively different in: thdir background characteristics. In -particular, the 4-*2 transfer.,students were more likely than withdrawals to have high SES scores., to include more Hispanic and students from theWest, and to have higher aptitude test scores and high educational aspiration. Even after some prior variables (stepdown analysis were, considered, the differences between the two groups of students on these variables (except high1 school program) were still significant.
Discriminmant analysis s.upported the, above findings even after all other variables were considered. However, as .shown by the sign of~the discrimiinantfunction coefficient, 4-t2 transfers tended to have lower college grades than withdrawals. The data seemed to suggest that. 47t2 transfers aspired more to obtain a college education than did withdrawals; thus, they enrolledin a 2-year college. if their performance was too poor. to cIontinue in a 4-year college.
Table 9-c.-- Test statistics for the comparison between 4->2 transfers and withdrawals
Standardized Variable ~~~~Univariate F' Stepdown F' discriminant
(d.f. =1, 3242) function coefficients3
SES .... ......... 4.82*4 .. 4.82* .. 0.23 Female (vs. male) . .. ........ 87 1.09 .25 Black (vs. noriblack).. ........ 05 .06 .14 Hispanic (vs. non-Hispanic).4.24* 5.9 6*.3 High school grades'........1.71 1.30 -0 Academic aptitude test scores. .... 754* .39.6.28*
Educational aspiration .l...... 0.54*1* 6.43* .33 Academic high school program
(vs. general and voctech) ....... 8.82* 3.30 3 College grades.......... l 2.47 -2 Academic field of study
(vs. nonacademic).. ....... 3.61 1.50 ..19 South,(vs. others)...........62 . .. 66 .27 West (vs. others).......8.72**.. 12.0 1** .. 54~
Multivariate F =3.85 X' 1 )=45.82 (d.f.=12, 3231).
pc .001~~~~~~p .0
NOTE.--I. Within-group variance is shown in table 8. 2. Variables are .listed in the order in which the stepdown analysis was performed. Thus, the stepdown
Fshows the significance of the indicated, dependent variable, controlling for all variables lsted above it.
3. The sign of the discriminant function coefficients shows the direction of relationship. A positive sign indicates that transfers were higher on the dependent variables than werIe withdrawals.
.4. *p < .O5; *pc<.01
B. Comparisons Between 2-Year Coliege Transfers and Nontransfers
The same techniques used in the comparisons of the 4-year college students were employed for the analyses of the 2-year college students. The weighted means and the pooled standard deviations on the selected background variables are presented in table 10, and the test statistics for group comparisons are included in tables 11-a to 11-d. The results were quite different from those of the 4-year college transfer and nontransfer comparisons.
SES........... 0.10 0.14 0.29 0.05 0.12 0.62 Female (Vs. male) ...... .43 .40 .45 .62 .52 .50 Black (vs. nonblack)...... .03 .04 .03 .03 .04 .17 Hfispanic (vs. non-Hispanic)..... .06 .09 .01 .02 .05 .19 Hfigh school grades...... 5.61 5.37 6.11 6.05 5.41 1.22 Academi'c aptitude test scores . . 53.35 51.55 55.02 54.31 52.38 6.20 Educational aspiration..... 5.04 5.20 5.38 4.57 4.6.95 Academic high school program
(vs. general and voctech) ... .54, .57 .67 .58 .142 .49 College grades .. ..... 5.40 6.02 5.92 5.31- 1.31 Academi'c field of study
(vs. nonacademic) .. South (vs. others) ..
... .
.22 .44 .42
West (vs. others)....... .43 .32 .24 .23 .31 .45 N3 . ..... 253 51 360 175 452 1291
Table 10.- Weighted means and standard deviations for various college-going groups on background variables (2-year colleges)
SES is a composite score with mean of 0, and standard deviation of 1. Aptitude test scores are standardized scores with mean of 50 and standard deivation of 10. High school and college grades were coded as follows: mostly A = 8; about halfA and half B = 7; mostly B = 6; about half B and halfC = 5; mostly C -_ 4; about halfC and halfD = 3; mostly D = 2;1 and mostly below D = 1. Educational aspirations were coded as follows: less than high school = 1; high school = 2; some voca- tional studies beyondhigh school = 3; two-year college = 4; four-year college = 5; and graduate school =- 6.
2The squares of thesevalues are within-group means of squares (the error terms for univariate analysis). The differences in sample size in this analysis and previous analyses Were due to missing data on background variables primarily because of nonparticipating in the base-year survey.
1. There were no differences in background variables between persisters and 2->2 transfers. The multivariate F-ratio of 1.07 was not significant at the .01 level'(see table I 1-a). The univariate F-ratios also failed to reveal any significant differences, and no significant discriminant function was* obtained. Thus, it was concluded that, at least on the selected variables in this study, those students who remain in a 2-year college and those who transfer to another 2-year college are not significantly different in their background variables.
26
Table li-a.-- Test statistics for the comparison between persisters and 2-+2 transfers
I ~~~~~~~~~~~Standardized Variable ~~~~~~Univariate F1 Stepdown F' discriminantI ~~~(d.f. 1, 1286) function
_________________________________________ j ~~~~~~~~coefficients.3
SES.........0.. .... 0.24 .2U.7 Female (vs. male) ..... .. .... 09 .08 .00. -Black (vs. nonblack)..... ...... 06 .1 03 Hispanic (vs. non-Hispanic) . . .. 7...1 .94 .29 High school grades....I. .. ... 1.61 .138 .-.09 Academic aptitude test scores ... ... 3.55 2.07 -. 54 Educational aspiration ........ 1.18 1.84 A45
Acadermic high school program (vs. general and voctech) --....... 16 .70 .22-
College grades................05 ~-.08 Academic field. of study
(vs. nonacademic) ..... ..... 1.37 2.14 -.45 South (vs. others) ..... .. .... 10 .17 ~.-.09 West (vs. others) ..... ...... 2.73 3.04 -.55;
Multivariate F =1.07,X( 2 12.76 (d.f.=12,4275) .
p < .38
NOTE.--I1. Within-group variance is shown in table 10. 2. Variables are listed in the order in which the stepdown analysis was performed. Thus, the stepdow'n
F shows the significance of the indicated dependent variable, controlling for all variables listed above it.
3. The sign of the discriminant function coefficients shows the direction of relationship. A positiv sign indicates that transfers were higher on the dependent variables than were persisters.
2. Students who moved to the 4-year colleges were, however, different from those who remained in the 2-year colleges (see table 11-b). The differences were significant on almost every individualvariable except sex and black-versus-nonbiack (the univariate F-ratios for those two variables were not significant at the .05 level). It can be seen from table 10 that 2->4 transfers had'a higher SES level, were composed of fewer'Hispanics,4had higher high school and college grades, and were more likely to major in academic fields than were those who remained in the 2-year college. The percentage of students in the West. who persisted in 2-year colleges was greater than the percentage of those who transferred to 4-year colleges. The opposite pattern held true for the South.
Some of these differences, however, became insignificant when some prior variables were held constant. As shown by the stepdown statistics, 2--4 transfers and persisters were similar in aptitude, aspiration, and high school program when SES, sex, and race were considered. The higher discriminantweights on high school* and college grades and academic field seem to indicate that 2->4.transfers may be a result of higher academic qualifications.
27
Table 11-b.--Test statistics for the comparison between 2->4 transfers and students who remained in 2-year colleges (i~e., persisters and 2->2 transfers)
Standardized Variable
1286). discriminant
function coefficients3
SES .............. 9.29** 9.29** .. 26 Female (vs. male) .......... ..66 .90 -.02 Black (vs. nonbiack).. ........ 38 .03 .01 Hispanic (vs. non-Hispanic)....... 12.84** 9.46** -.24 High school grades.......... 29.34** 31.20** .37 Academic aptitude test scores...... 19.7 1* 2.28.0 Educational Aspiration.. ...... 8.92** 3.60 .14 Academic high school program
(vs. general and voctech) .6.8.......82 .02 College grades.... ....... 34.57** 12.52** .48 Academic field of study
(vs. nonacadernic) .. ....... 16.99** . 10.00**' .33 South (vs. others)......... 5.77*' 4.67* .12 West (vs. others) .. ........ 10.48** 7.08* -.32
Multivariate F = 7.85 X2 (1 2 ) = 91.23 (dif. = 12, 1275) p<.0
p < .001 p<.0
NOTE.-- 1. Within-group variance is shown in table 10. 2. Variables are listed in the order in which the stepdown analysis was performed. Thus, the stepdown
F shows the significance of the indicated dependent variable, controlling for all variables listed abok'e it.
*3. The sign of the discrinminant function coefficients shows the direction of relationship. A positive sign indicates that transfers were higher on the dependent variables than were students who re- mained in the 2-year college.
*4. *p< .O5; **p< .o1
3. The 2->4 transfers were also different from withdrawals and graduates (i.e., students who completed a 2-year program but discontinued further study) in their background characteristics. The differences are shown in tables 11I-c and 11I -d. In particular, the 2->4 transfers were higher than withdrawals in SES, academic achieve- ment, educational aspiration, and field of study.
The comparisons between 2-4 transfers and graduates revealed some interesting information. Graduates were more likely than 2-*4 transfers to be female students, and they scored lower on SES and educational aspirations (see table I1l-d). However, there were no significant differences in ability and achievement scores between graduates and 2-4 transfers. The greater proportion of graduates in nonacademic programs seems to indicate that most graduates considered the 2-year college education as their educational goal. This group of 2-year graduates contained more female and lower SES students than did the 2-N transfer group.
28
Table 11-c.--Test statistics for the comparson between 2-÷4 transfers and withdrawals
Standarded Univari'ate F' Stepdown F2 discriminant
Variable (d.f. = 1, 1286) function coefficients3
SES ......... ..... 16.05 **4 16.05** 0.16 Female (vs. male).3.44.......2.76 -.17 Black (vs. nonblack)..........1.23 .16 .00 H-ispanic (vs. non-Hfispanic).......6.93* 3.89 -.09 ffigh school grades.......... 64.65** 81.22** .37 Academic aptitude test scores ...... 36.22*~* 3.07 -1 Educational aspiration.. ... I.... 85.99** 52.46** .37 Academic high school program
(vs. general and voctech) .. ..... 53.14** 18.54** .24 College grades ........... 59.23** 19.12** .38 Academic field of study
(vs. nonacademic)......... 88.12** 37.89** *.42 South (vs. others)..........6.26* 3.82 .10 West (vs. others) ... ....... 5.42* 1.81 -.10
Multivariate F =21.54 (12 - 236.24 (d.f. = 12, 1275) p<.0
p < .001~ ~~~p< 00
NOTE.--l. Within-group variance is shown in table 1 0. 2. Variables are listed in the order in which the stepdown analysis was performed. Thus, the stepdown
F shows the significance of the indicated dependent variable, controlling for all variables listed' above it.
3. The sign of the discrinminant function coefficients shows the direction of relationship. A positive sign indicates that transfers were higher on the dependent variables than were withdrawals.
4. *p <.05; **p< .o1
Table 11-d.--Test statistics for the comparison between 2-'4 transfers and graduates
Standardized Univariate F' Stepdown F2 discriminant
Variable (d.f. =1, 1286) function coefficients
SES............... 18.32* *4 18.32** 0.23 Female (vs' male),. ........ 12.48** 11.92** -.19 Black (vs. nonblack.)..........00 .76 .03 Hlispanic (vs. non-Hfispanic.. ...... 02 .46 .03 Hfigh school grades.. ......... 22 3.29 -.01 Academic aptitude test scores ...... 1.54 .07 -.12 Educational aspiration:.. .... I86.56** 69.40** .55 Acadermic high school program
(vs. general and voctech).......4.32* .19 .02 College grades............69 1.22 .16 Academic field of study
(vs. nonacademic).. ....... 91.66** 47.50** .58 South (vs. others)..........7.69* 3.99 .21 West (vs. others) .. ......... 01 1.30 .11
Multivariate F =13.73 -=155 (d.f.= 12, 1275). X'12)=152
p<.001. ~~~~~~p< .0 0 l
NOTE.--I. Within-group variance is shown in table 10. 2. Variables are listed in the order in which the stepdown analysis was performed. Thus, the stepdown
F shows the significance of the indicated dependent variable, controlling for all variables listed above it.
3. The sign of the discrimfnant function coefficients shows the direction of relationship. A positive sign indicates that transfers were higher on dependent variables than were graduates.
4. *p <.05; **p <.01 2
C. Summary and Discussion
Transfer students were, in general, different from nontransfer students. In 4-year institutions, 4-+4 transfers tended to have higher levels on SES and college achievement but lower aptitude, and tended to include more female students than did persisters. This seems to suggest that those studentsmoving among the 4-year colleges were students who, had the qualifications for greater mobility-high SES background which reduces financial pressure, and high achievement which would'be accepted by other colleges. However, why there were more female than male 4-*4 transfers is unknown. It might be that female students have more difficulty than do males in finding a suitable opportunity for career development or a satisfactorysocial life on campus.
Transferring to a 2-year college after .2 years of study in a 4-year college was an unexpected phenomenon. The generally lower grades of those' 4-~2 transfers may indicate that they may have had academic difficulties in the 4-year institutions. However, the data showed that the 4-*-2 transfers had high educational aspiration; perhaps* they* intended to improve their achievement in a 2-year college and then retumn to a 4-year college, (e.g., Kuznik, 1972) or at least get a 2-year college degree that might be helpful in career development. The future NLS survey will provide data for testing this assumption. The 4-~2 transfers' higher SEIS background and higher aspiration were probably the underlying factors that contributed to their desire to continue their edu- cation rather than to withdraw. entirely.
In 2-year colleges, 2-~2 transfers were not significantly different from persisters. However, 2-~4 transfers were a distinctive group among the 2-year college students; they 'had higher scores on SES and achivemnent, and they were more likely to major in the field of academic studies than were other groups of students. A number of reasons might explain why these students transferred to a 4-year school. Many of the 2->4 transfers no doubt were.students who aspired to a 4-year college education, but such reasons as inadequate prpeparation in high school or inadequate academic qualifications led: them to enroll in a 2-year college initially. There might be some financial. considerations involved since 2-year colleges are generally less expensive than 4-year colleges. There might also be some decision problems. Many students may not know what they want to study or what they want to do in the future;, thus, they simply use a 2-year college as a way station until their goals are set. The comparisons between this group of students and students enrolled in the 4-year colle~ge immediately after high school graduation will be informativel. Some comparisons are included in the next chapter.
30
IV. COMPARISONS BETWEEN 2- 4 TRANSFERS AND 4-YEAR COLLEGE NATIVE STUDENTS
Going to a 2-year college initially and then transferring to a 4-year college, rather than enrolling in a 4-year college immediately after high school, is considered by many students as a satisfactory program of,,higher education. A recent study by the, Carnegie Commission of Higher Education (1970) revealed that over two- thirds of the students entering 2-year colleges intended to transfer to 4-year colleges. The NLS data, as pre- sented, in chapter II, showed that about a quarter of the 2-year college students, did transfer to: 4-year colleges within two years. Therefore, it is informative to examine the characteristics of the students taking these two alternate paths, and to compare them as to financial aid status, academic performance, and satisfaction with education.
A. Comparisons on Background Variables and Individual Characteristics
The' first question addressed is whether the choice of different college-going pathi i's related to the students' backgrounds anid/or certain personal characteristics. To answer ths uestiotn the 2->4 transfers anid 4-year college-native students' were compared on the followingvariables:
(1) Background characteristics: sex, race, and socioeconomicbackground; (2) Region where the student graduated;
()High school programs; (4) Academic performance: high school grades, aptitude test scores; (5) Educational aspiration; (6) Self-concept and locus of control; (7) Life goals: work, community, and family-oriented life goals. The Variables of self-concept, locus of control, and life goals were psychometrically-constructed scales,
measured when the students were seniors in high school. They were included on the assumption that they might influence an individual's choice of different educational or career paths. The scale definitions are presented in appendix D. Both self-concept and locus of control were measured on a 5-poIint scale. A high score on locus of control indicated a high degree of internality; a low score, a high degree of externality. A high score on self- concept indicated positive self-concept;' Life goals were composites based upon items with a 3-point scale, ranging from not important (1), to very important (3). Other selected variables,, such as SES and educational aspiration, were specified in* the preceding chapter; the'same definitions were applicable to the analysis in this chapter.
The weighted means of common standard deviations on the selected variables are' presented in table 12. The test statistics (F-ratios) of the group differences are included in table 13. As expected, these two groups of students differed in their backgrounds and characteristics. (The multivariate F-ratio of 19.00 is significant at the .001 level with degrees of freedom of 15 and 2792). The univariate F-ratios in table 13 show that-native and transfer students in 4-year colleges differed significantly on most of the selected variables. Native students tended to have higher SES scores, high school grades, aptitude tests, and educational aspiration than did transfer students. This finding was consistent with previous findings (e.g., Kintzer, 1973; Brinbaum, 1970). Native stu- dents were more likely than were transfer students to have been graduated from high-school academic programs, to have higher self-concepts, and to be more internal in locus of control. On the other hand, transfer studerits had higher scores on work-oriented life goals than native students, and were composed of proportionallymore ,nonblack students. In addition, there were proportionally more transfers than native students in the West than in other regions.
1Native students were those students who attended 4-year colleges immediately after high school graduation and who, after two years, persisted. in 4-year colleges or who transferred to other 4-year colleges.
3 1
Means ~Common Variable Masstandard
Native Transfer deviation
1. SES........... 0.47 . 0.29 0.69 2. Female (vs. male)........... .... 49 .45 .50 3. Black (vs. nonblack)............. .05 .03 .22 4. Hispanic (vs., non-Hispanic) ..... .. .... 01 .01 .12 5. Highschoolgrades ............. 6.66. 6.11 1.14 6. Aptitude test..... .. .... 58.06 54.80 5.66 7. Educational aspiration..........5.68 5.39 .56 8. Academic high school program.
(vs. nonacademnic)... .84 .. 68.3 9. Self-esteem................. 4.02 3.91 .64
10. Locus of control............... 4.07 3.9